Optical coherence tomography (OCT) systems with higher imaging speed have a significant diagnostic value, since they allow increased field of view, reduced measurement time, improved resolution down to the cellular level by reducing motion artifacts, and supporting functional imaging such as OCT angiography. Therefore, increasing the imaging speed without increasing the burden of higher costs is one of the ultimate goals of the next generation OCT systems. While ultra-fast OCT systems with A-scan rates exceeding tens of MHz have been demonstrated, a major roadblock in widespread clinical application of such systems is their significant cost due to their reliance on ultra-fast lasers and detectors. Parallel SS-OCT systems have long been considered the best solution to reduce the cost of ultra-fast OCT systems, since both the laser swept rate and detector bandwidth can be scaled down by increasing the number of parallel channels. However, the need for balanced detection for achieving high sensitivity has limited the number of parallel channels to a handful, and rendering such systems unattractive. The main limitation is the balanced detection, which requires stringent tolerances for the equalization of the power across each detector pair. Also, achieving shot-noise limited sensitivity in this method requires a large reference power per detector that could add up, as the number of channels grows, to a prohibitively large laser power. Mohseni's research group recently published results based on a new detector called Electron Injection (EI) showed that shot-noise limited sensitivity of ~105 dB could be achieved without balanced detection for the first time. Additionally, the required reference power for shot-noise limited performance was about three orders of magnitude lower than the conventional detectors. Based on these findings, this project hypothesize that a parallel SS-OCT system with shot-noise limited sensitivity can be made using the new EI detector; and that the parallel OCT is highly scalable and can be augmented to Multi-MHz scan rate by simply adding low-speed channels and using a conventional low-speed swept source. This hypothesis will be addressed in the experiments organized in the following Specific Aims: (1) to build a parallel OCT system with 16 parallel channels based on EI detectors, and operating at the shot-noise limited sensitivity. (2) to scale up the system to 64 parallel channels, each operating at 100 MHz, leading to 3.2 Mega A-scans per second, while using a conventional swept-source at 50 KHz swept rate and requiring only a few mW of laser power. Should this exploratory study achieve the planned goals, it paves the way for both low-cost ultra-fast OCT systems and also OCT systems with potentially unprecedented scanning speeds. The team members have made large arrays of EI detectors with thousands of elements in the past and are familiar with the industrial and clinical limitations. If success, this project enables parallel SS-OCT with over 1000 parallel channels in the future, and unprecedented speeds approaching 100 Giga-voxels/second.